diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Multi-Way (http://www.diyaudio.com/forums/multi-way/)
-   -   JBL 2445J HF boost with autoformer in reverse mode (http://www.diyaudio.com/forums/multi-way/242408-jbl-2445j-hf-boost-autoformer-reverse-mode.html)

elac310 12th September 2013 05:10 PM

JBL 2445J HF boost with autoformer in reverse mode
 
Hello all,

Had an opportunity to get a set of blown JBL 2445J, with 2382a horns.
I had some NOS JBL diaphragms laying around (D16R2445) and decided I'd give it a try in my system, as the 3rd way covering 2000 Hz upwards. I had often heard that 2" drivers are capable of reproducing HF sounds in a way no 1" driver can. Needed to check that...

Drivers were cleaned for possible residues in the voice coil gap (they were clean), new diaphragms installed - very easy with the centering structure carved in the pole plate and first quick listening test: surprise, surprise, there is life in the UHF region, clearly more than what I was used to from earlier tests with a (much more recent) 2" driver.

Tested in my audio-room, about 35m^2:
At 50 cm, response graph is almost linear up to 13-14 kHz before it drops.
At 5 meters, the drop starts at 10 kHz.

I first tried a standard 6dB xover, with a HF boost circuit, used in vintage and cinema crossovers, consisting of a 3uF capa + 0,02mH coil and was not impressed by the results; in fact, hearing no difference, I concluded I had done something wrong.

Next, I used a simple 6dB filter of 3.3uF, with in parallel, an autoformer with multiple -1,5 db taps, but mounted in reverse mode i.e. increasing the signal at the expense of impedance drop, followed by a small cap of 1 uF. I applied +6 and then +7,5 dB and I think I've got something which makes any tweeter totally superfluous. And what an impact on cymbals, bells and many other instrumenst which have become so much real now !

But I Need to do some further tests to track possible distorsion and watch for excessive impedance drop on the amplifier side.
Measurements also need to be done.

ra7 12th September 2013 05:22 PM

Above 6 kHz or so, there is a lot of breakup and it is easily seen in the FR. You can extend the bandwidth by raising the response, but you are not going to see a smooth response.

elac310 13th September 2013 02:06 AM

I'm really positively surprised by the overall performance of the 2445/2382 combo.

Can't say I have heard many anomalies so far (with or without boost); in fact, most seem for the time being to be connected to my approximative crossover set-up and to the recording quality. I discover more and more of these (overdriven mikes and mike amplifiers, imbalance between channels, excessive brillance added to female voices etc.).

A few minutes ago, I just listened to Alela Diane's "Pirates' Gospel" which is too brightly/metallically recorded and thus a good test I guess, and I found on track 11 a series of distorsions on her voice which I had never heard - or detected ? - until now. Will switch back to another driver tomorrow to find out.

I'm fully aware of what others have already said/observed about limits of 2" drivers, titanium+non-ribbed/non-aquaplased diaphs, old JBL designs as 2445 could be etc. But want to see this by myself, and to find out also how the extra boost impacts negatively and audibly on the result.

To my taste, the main recurring issue is in fact the horn, I think, which seems to radiate too broadly in my room at that listening distance, and that affects clearly the stereo-image (central image is too big).

weltersys 13th September 2013 08:21 AM

Quote:

Originally Posted by ra7 (Post 3629499)
Above 6 kHz or so, there is a lot of breakup and it is easily seen in the FR. You can extend the bandwidth by raising the response, but you are not going to see a smooth response.

And the 2445 breakup in the upper end response can also be seen in the distortion response on the 23XX series horns, as I recall a 6 kHz tone resulted in a 12 kHz tone near equal in level at relatively low SPL.
"Sparkly" high end, but not "high fi".

elac310 13th September 2013 10:48 AM

Thanks for feedback.
"Sparkling" is OK for me as long as it sounds natural and genuine, i.e. it's what I hear at the concert or it's what the sound engineer wanted when doing the final mix/master record (e.g. Rudy Van Gelder on Oscar Peterson's track "you look good to me", album "We get requests", where the level of HF/the cymbal is progressively increased).

But agree with you and I've myself been chocked by the number of (commercial or not) speakers which add too much sparkle, typically a compression tweeter coupled to average mid/woofers of 7": even if on paper response is straight at 1m, subjectively I often find the tweeter overly present due to its higher energy/impact etc. Also once met a former DJ and PA supplier, who had invited me to his home. There were 4 compression tweeters on top of his living room speakers and it did hurt my ears (his wife also left immediately the room). I drew politely his attention to the fact that there was way too much trebble but that's the way he loved it he said.

I love in a small jazz pub the impact/energy/sound of a drum set, a cymbal, a double bass for instance. I also have regularly the chance of listening at very close range to 50-60 head-choirs, singing for instance Beethoven's 5th. It's something absolutely unique. Bringing all this into my music room, that's what turns me on and why I spend time and money on this hobby.

moray james 13th September 2013 06:29 PM

1 Attachment(s)
Quote:

Originally Posted by elac310 (Post 3629474)
Hello all,

Had an opportunity to get a set of blown JBL 2445J, with 2382a horns.
I had some NOS JBL diaphragms laying around (D16R2445) and decided I'd give it a try in my system, as the 3rd way covering 2000 Hz upwards. I had often heard that 2" drivers are capable of reproducing HF sounds in a way no 1" driver can. Needed to check that...

Drivers were cleaned for possible residues in the voice coil gap (they were clean), new diaphragms installed - very easy with the centering structure carved in the pole plate and first quick listening test: surprise, surprise, there is life in the UHF region, clearly more than what I was used to from earlier tests with a (much more recent) 2" driver.

Tested in my audio-room, about 35m^2:
At 50 cm, response graph is almost linear up to 13-14 kHz before it drops.
At 5 meters, the drop starts at 10 kHz.

I first tried a standard 6dB xover, with a HF boost circuit, used in vintage and cinema crossovers, consisting of a 3uF capa + 0,02mH coil and was not impressed by the results; in fact, hearing no difference, I concluded I had done something wrong.

Next, I used a simple 6dB filter of 3.3uF, with in parallel, an autoformer with multiple -1,5 db taps, but mounted in reverse mode i.e. increasing the signal at the expense of impedance drop, followed by a small cap of 1 uF. I applied +6 and then +7,5 dB and I think I've got something which makes any tweeter totally superfluous. And what an impact on cymbals, bells and many other instrumenst which have become so much real now !

But I Need to do some further tests to track possible distorsion and watch for excessive impedance drop on the amplifier side.
Measurements also need to be done.


elac310 14th September 2013 12:15 PM

@Moray James

I guess you're referring to these Klipsch crossovers because they make use of autoformers in the HF path.

Do you know whether these were used not to attenuate but to raise the HF level ? At a first sight, the fact that the cap before the autoformer (between the latter and the amp) is of relatively low value compared to the 15uF after the transformer suggests that it was used classically, as an attenuator.

Another difference is that they use the auformer for the whole of the HF path, whereas in my case, the autormer is part of a circuit in parallel to the main HF path, to raise just the VHF.

(+)AMP=============3.3uF====================(+)244 5J

...............................auto-transf.
................................originally IN==============1uF===(+)2445J
..............................or. -1.5dB tap, now (+1.5)
.................................or. -3dB tap, now (+3)
.................................................. .........(+4.5)
.................................................. .........(+6)
(+)AMP=======100uF===========(+7.5)
.................................................. ..........(+9)
.................................................. ..........(+10.5)
.................................................. ..........(+ etc.)
.................................................. .........(mass connect)

(-)AMP=====================================(-)2445J

elac310 6th October 2013 06:52 PM

Update:

I've now gone for the crossover cap before the autoformer. As a result, the 1uF has been removed, the 100uF is replaced by a 6.8 (or 5.6) uF.

The most intriguing discovery for me was caused by the recent swap of the 3.3uF on the main branch, with a USSR vintage NOS capacitor said to be PIO, in lieu of a standard MKP. Although I've spent almost 30 years in building speakers and experimenting with different technologies of components in Xovers (I had reached the conclusion that variations in sound quality are marginal when Xover is properly designed), I was really surprised by the positive changes on the sound induced by these capacitors (smooth yet detailed, delicacy, timbre etc.). The 2445 with TI diaphragms (in new condition) is very revealing so it could be that the sonic signature of X-over componentry becomes particularly audible (I also found out a test wire with croco was bad). "Oh my Mama" sung by Alela Diane does not sound distorted anymore, idem with Natalie Desay on Offenbach's Hofmann Tales "les oiseaux dans la charmille" (Erato). Violines and many other acoustical instruments have also a very pleasant presentation now.

Provisional conclusion: 2445 can sound smooth even with TI diaphragms, provided the right componentry is used.

elac310 6th October 2013 06:57 PM

Ah, and one more thing: still no extra tweeter needed at all so far. The HF boost brings the extra sparkle that is missing originally. A good test is Livingston Taylor's "Ink" CD (Chesky), with the little bells and other UHF sonics in the background.

moray james 6th October 2013 09:09 PM

the auto transformer in the CF3&4 are used to stet up the signal to boos HF response. More than that I cannot say. Best regards Moray James.


All times are GMT. The time now is 03:37 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2